[Feature Request] Rich Logging Option

[natolambert]

🚀 Feature Request

A configuration option to save all data for every trial.

Motivation

When debugging it can be very useful to see the intricate details of how different trials played out, without having to re-run computationally intensive experiments.

Is your feature request related to a problem? Please describe.
The replay buffer class is very efficient and elegant for running the algorithms, but saving more information in a more structured manner could be useful.

Pitch

Describe the solution you'd like
Core things to be logged in this mode:

• raw trajectory data of each trial,
• dynamics model of each trial,
• any optimizer parameters of each trial that may change.

Advanced things to be logged in this mode:

• planned action sequences from steps in a trial,
• some sort of snapshot of the training and validation data at each trial.

Describe alternatives you've considered
The trajectory buffer is a step towards this direction.

Are you willing to open a pull request? (See CONTRIBUTING)

Additional context

Add any other context or screenshots about the feature request here.